Relational analytics for the CRE broker or CMBS buyer
APIP — Autonomous Property Intelligence
Methodology

How APIP builds the canonical property record.

HI → AI = PI · Human Intelligence into AI equals Property Intelligence

Every record is assembled from 14 source classes, conflict-resolved by a deterministic hierarchy, scored for completeness and recency, and compliance-screened before delivery. This page documents every step.

Conflict resolution · live
Field · owner_legal_name
3 sources, 3 values
  • FL SUNBIZGOVERNMENT
    WINNER
    PROVISO ENTERPRISES, LLC
  • LEE CLERK · DEEDGOVERNMENT
    Pulled 2h ago
    PROVISO ENTERPRISES LLC
    Discarded · Rule 3 · cross-source agreement
  • ENRICHMENT T1COMMERCIAL
    Pulled 12d ago
    Proviso Enterprises
    Discarded · Rule 2 · source authority
Resolution

Two sources agreed on entity name (Rule 3). The agreed value comes from the government authority with a 6h-old pull (Rule 2 + Rule 1). Commercial enrichment value held as alternate, never overwritten.

Real production resolution · audit log on request
0
Source classes
Government + enrichment + feedback
0
Fields per record
Owner, debt, compliance, contact
0
Conflict-resolution rules
Deterministic, auditable
0%
Source-stamped
Every field, every record
Source hierarchy

14 source classes, each with documented authority and refresh cadence.

Every field in every APIP record carries a source ID, pull timestamp, and confidence tier. The table below documents 8 of 14 source classes. Full list and authority weights available on request.

IDSourceTierRefreshAuthorityConf.
CR-01
County Assessor / Property Appraiser
Parcel ID, legal description, assessed value, land area, building SF, year built, owner name
PRIMARY30–90 daysGovernmentHIGH
DR-02
Deed Registry / Clerk of Courts
Grantor, grantee, deed type, consideration, book/page, recording date, instrument number
PRIMARY24–72 hoursGovernmentHIGH
SB-03
SunBiz / Secretary of State
Entity name, type, status, registered agent, manager, formation date, EIN
PRIMARY24–48 hoursGovernmentHIGH
CS-04
CMBS Servicer Data
Loan amount, maturity, LTV, DSCR, IO period, servicer, watch-list status
PRIMARY< 24 hoursServicerHIGH
DNC-05
National DNC Registry
DNC status, registration date, litigator flag, TCPA cell classification
COMPLIANCE14 daysFTCHIGH
EN-06
Paid Enrichment (Tier 1)
Phone numbers, email addresses, mailing address, contact role
ENRICHMENT30 daysCommercialMED
EN-07
Paid Enrichment (Tier 2)
Social profiles, business relationships, additional contact methods
ENRICHMENT60 daysCommercialMED
CRM-08
CRM Outcome Feedback
Dial outcomes, corrections, closed deals, contact quality ratings
FEEDBACKReal-timeOperatorMED
Confidence scoring

Three dimensions, one defensible confidence tier.

Every field receives a composite score across completeness, accuracy, and recency. The score maps to a HIGH / MED / LOW tier that tells you exactly how much to trust the value, and what validation is recommended.

Completeness

30%

Is the field populated? Is the value non-null, non-placeholder, and within expected range?

Accuracy

40%

Does the value agree with at least one independent source? Has it been validated against business rules?

Recency

30%

How old is the source pull? Recency decay applies: government records decay slower than enrichment data.

HIGH≥ 0.80

Cross-source agreement, recent pull, government authority. Safe to act on without manual validation.

MED0.60 – 0.79

Single-source or older pull. Recommend spot-check before high-stakes outreach.

LOW< 0.60

Inferred, stale, or single-enrichment-source data. Manual validation required before use.

Conflict resolution

When sources disagree, the hierarchy decides.

APIP applies a five-rule deterministic hierarchy when sources conflict. Every conflict resolution is logged with the winning source, the losing source, and the rule applied, available on request.

  • Deterministic, no black-box ML
  • Auditable, full conflict log on request
  • Operator-correctable, CRM feedback applies immediately
  • Documented, every rule is published here
Conflict resolution hierarchy5 rules
  1. 1

    Recency

    The most recently pulled value wins, subject to source authority weighting.

  2. 2

    Source authority

    Government records (county, deed, SunBiz) outrank commercial enrichment for the same field.

  3. 3

    Cross-source agreement

    If two independent sources agree, the agreed value wins regardless of recency.

  4. 4

    Business-rule validation

    Values outside expected ranges (e.g., negative LTV, future formation date) are flagged and held pending review.

  5. 5

    Operator override

    CRM-sourced corrections from verified operators apply immediately and are logged with source attribution.

Refresh cadence

Every source on a documented refresh schedule.

Refresh cadence is determined by source authority, data volatility, and compliance requirements. Government records refresh on filing events. Compliance screens refresh on a fixed 14-day cycle.

CMBS Servicer Data
Trigger: Servicer file drop
< 24 hours
Deed Registry
Trigger: Recording event
24–72 hours
SunBiz / SOS
Trigger: Filing event
24–48 hours
County Assessor
Trigger: Assessment cycle
30–90 days
DNC Registry
Trigger: Scheduled re-screen
14 days
Paid Enrichment T1
Trigger: Scheduled refresh
30 days
Paid Enrichment T2
Trigger: Scheduled refresh
60 days
CRM Feedback
Trigger: Operator dial outcome
Real-time
Intellectual property

5 patents pending. The method is the moat.

APIP has filed provisional patent applications covering the core mechanisms behind owner motivation scoring, entity piercing, and lead-time prediction. The inventions listed below are the algorithms that produce the ranking, not interface features.

5 Provisional Applications Pending, USPTO
1
Patent Pending · Claims 1-16

Self-Calibrating Lead-Time Engine

Reduced to practice

Dynamically adjusts the predicted sell window based on validated transaction outcomes. The engine recalibrates without manual input each time a scored lead results in a confirmed sale.

2
Patent Pending · Claims 17-54

Owner-Motivation Composite with Entity Piercing and Cross-Portfolio Distress

Reduced to practice (core); enabling description (Sec G)

Multi-signal composite scoring that pierces LLC, trust, and partnership structures to the human principal, then applies financial distress signals across every property that principal controls.

3
Patent Pending · Claims 32-36 (continuation)

Environmental-Event Proximity Model

Specification stage

Geospatial model that measures the proximity of environmental events (code violations, permits, adjacent actions) to a parcel and converts proximity-time relationships into motivation signals.

4
Patent Pending · Per PROVISIONAL-4

No-Leakage Point-in-Time Validation

Reduced to practice

Pipeline architecture that constrains each historical scoring snapshot to data available at that exact point in time. Prevents future knowledge from contaminating past-period accuracy measurements.

5
Patent Pending · Claims 37-45

County Market Formation Factor

Reduced to practice

A per-county contextual multiplier derived from Census Bureau Business Formation Statistics that adjusts individual parcel scores for local market acceleration or contraction at run time.

Disclosure note: Patent applications are provisional filings. Final claim scope is subject to USPTO examination. Nothing on this page constitutes a representation of granted rights. Contact craig@nuro.is for licensing inquiries.
Questions about the methodology?

We publish our methodology because we stand behind it.

If you have questions about a specific source class, conflict resolution rule, or confidence tier definition, contact us. We'll respond with a documented answer, not a sales call.